The chances of avoiding cervical cancer are improved by early detection of abnormal cells in the cervix. Automated methods are developed because manual detection is time-consuming and error-prone. This work presents a comprehensive analysis of AI-based methods for cervical precancerous detection, screening, and prognosis. From 2538 initial publications, 117 eligible studies were analyzed. AI systems can distinguish benign from malignant cervical cytology with 80-100% accuracy and predict CIN2+ with 71.9-98% sensitivity and 51.22-96.2% specificity. AI can supplement human judgment in interpreting cervical smears and images, especially where access to specialized facilities is limited. © 2025 Elsevier B.V., All rights reserved.